Post By JRIowa
Post By dinosaur
Post By mfgbydesign
Large lathe accuracy specs
What is a reasonable Z and X linear accuracy and repeatability specification to ask for?
I've been tasked with finding a rebuilder for a large horizontal lathe. 1400mm X 9000mm (55 inch swing, 35 feet of ways).
Machine is a Gurutzpe B1800/3.
I'm writing a request for proposal to machine rebuilders and the lathe doesn't have any specs on linear accuracy, but I'm asking for a different drive system for Z to replace the single pinion rack. Possibly a dual pinion or rotating nut ball screw. Scale feedback.
Last edited by mfgbydesign; 12-07-2012 at 11:33 AM.
Reason: put the question first
There is an ISO standard for machine tool acceptance. I've looked in my files and can't find it. Cross slide should probably be in the neighborhood of 0.0001"/foot for square and straight. For Z, I'd go with a scale (I'm a Newell guy). A DSM/DSG scale can go up to about 40'. You'd have to call them.
Don't forget all of the other specs that need to be written into your rebuild request. OSHA CFR-1910.212, 1910.219, 1910.219, ANSI Z-535, and applicable ANSI B11 stuff (it's B11-6 for manual lathes and B11.22 for CNCs). Almost forgot the NFPA 79 requirements.
For the rebuild, either B&O or Phoenix
Cut-Off Lathes, Rotating Head Cut-Off, Double-End Face, Chamfer and CNC Turning Machines - Bardons & Oliver
Welcome to Phoenix, Inc.
If you send me a PM with the name of the company you work for, I'll send you a copy of my rebuild specs for VTLs which should get you close.
Thanks for taking the time to reply.
The machine was originally accepted on DIN 8607 Items 1-15 which cover the alignment of the axis to the spindle, but not linear accuracy.
I have a copy of ASME B5.57 which describes various perfomance tests. Here is a couple line items in my RFP that I'm trying to fill in the blanks:
5.Test per: ASME B5.57 Section 7.2 Positioning Accuracy and Repeatability, Linear Axes
a.18.104.22.168 Bi-directional Accuracy of Positioning of an Axis (Z & X) equal or better than _____
b.22.214.171.124 Bi-directional Repeatability for an Axis (Z & X) equal or better than _____
6.Test per DIN 8607 Items 1-15 to be witnessed by customer after installation
7.Bi-directional Yaw (shake) not to exceed _____ to be verified at intervals not to exceed 12"
The original specs for the machine make no claims of linear accuracy, and the machine has a "shake" problem that I'm trying to correct. Shake is when a reversal in one axis causes a change in the other axis. In this case, reversing Z makes the tool move in or out in X, making it very difficult to hold <.001 tolerance diameters.
The machine has scales now which will be replaced. It has a rack drive on Z which has excess lost motion, so I am having this replaced with a different solution as well.
CNC lathes of large capacity :: Tornos Gurutzpe ::
FWIW You can't make a silk purse out of a sows ear, and I've seen a lot of money spent tryting to So I would contact the builders and see if they've the original specs (which I would think a co that builds that sort of machinery would have on file)
How true, the "tooling engineer" I'm working with wants the 1989 milling attachment to do rigid tapping. Yeah and I want sharks with friggin laser beams on their heads!
Originally Posted by Limy Sami
You can spend some money on the machine and probably make it better that it was. How good do you need it. Just as Sami said, it's not a silk purse.
a & b are going to depend on scale or whatever you put on. The shake could be a lot of problems like loose gibs or cross slide not square with bed.
Like a lot of machines the ways are worn where most of the work has been done. The gibs are tight in the unworn section and move to the middle the carraige floats around. The ways are going to be ground. I don't want to be unreasonable but need something to hold the supplier to. I hate to but may start browsing other machine specs to see what they claim. There's nothing on this from the OEM other than the alignments. The machine couldn't hold tolerance when it was new, we can't machine our shafts right now, and are constantly on the hunt for new vendors because nobody wants to do it twice. If you've got a large lathe in the midwest & want some real headaches PM me with size & I'll hook you up.
On the silk purses / sows ear question I wonder if there is any cost effective add on system which could actively correct for any residual errors in alignments. Motion Guru would probably know the up to date state of the art but back in the 1970's I did encounter some references to work on a similar problem of machine tool alignment when investigating the performance of the then new continuous position sensing (posicon) silicon detector relative to conventional 4 element quadrant detector in optical CLOS systems.
As I recall things the reports claimed that laser based active error compensation could compensate machine alignment variations of the order of tenths of an inch leaving residual tool-path errors of under a thou over tens of feet. The posicon based system was said to be easier to set up, less complex optically and far more tolerant in use. I was able to confirm the operational and installation advantages of the posicon and demonstrate error measurements of a thou or so over a 40 ft lab with both laser and (later) projector bulb sources. About £50 to £100 worth of gubbins at the detector end for the posicon, lots more for a quadrant. If you have CNC control it should be possible to inject the correction signal form that sort of system direct into the control loop. For a manual machine something akin to a hydraulic tracer set-up should be do-able. An obvious way would be to dump the screw cross feed and drive the laser source from side to side so the hydraulics do all the work. $64,000 question is whether there is anything tried and tested on those lines out there and ready to fit.
I imagine the machine tool research work in the reports I saw was aimed at aircraft wing skin milling and similar duties. Think there were also some side references to work on active control of gib contact forces for maximum stability. Research reports only of course so not a clue if anything real happened.
I have been involved in precision machine tool repair, rebuilding and manufacturing for 35 years.
My expertise is restoring the accuracy to worn out machine tools.
The relevant standards are as follows:
There is a paper published in October 2007 by Professor Brigid Mullany of the University of North Carolina in Charlotte.
She compares the various standards.
It is painfully entitled:
"Evaluation and Comparison of the Different Standards used to define the Positional Accuracy and Repeatability of Numerically Controlled Machining Center Axes"
I don't have the direct link, but you can google it or go to the AMT website and that has a link to it.
Here are the questions that you need to have the answers to:
1) What do you wish to accomplish with the machine after it is rebuilt?
2) Will it be dedicated to a particular part or a family of parts or general work?
3) If it is dedicated to a limited application, what material will you be cutting?
4) How many parts do you need to make in a given time period?
5) Will the machine be part of a cell or stand alone?
6) What are the most critical dimensions that you must hold?
7) What industry are the parts intended for?
8) What has the machine been used for in the past?
9) What are your likes and dislikes of the machine in its present state?
10) What sort of foundation will the machine be on?
11) What sort of environment is your factory? (Like a coal mine or a semiconductor clean room?)
12) What other types of machinery will be in the near vicinity?
13) Will there be special coolant used?
14) Will there be special chip handling requirements.
15) How will the rigging of the machine be handled if it shipped out of your plant?
Things to avoid:
1) Unrealistic tolerances will only increase the cost substantially.
2) Unrealistic time frames for the work to be accomplished. You will only attract rebuilders desperate for the work.
3) Changing your specifications in the middle of the rebuild.
4) Insisting on a certain way of accomplishing the goals without being open minded to alternatives.
5) Spending too much money on control capabilities that are unnecessary.
6) Skimping on the mechanical side of the project.
7) Not mounting the machine on a proper foundation.
8) Not investing in training the people who will have to run the machine.
9) Relying too much on software for compensation instead of proper mechanics.
10) Forgetting to consider the serviceability of the machine.
11) Not having your maintenance people involved in the project from the beginning.
12) Saving money with amateur riggers.
13) Not getting proper documentation such as electrical schematics, ladder diagrams, alarm codes, parameters, etc.
14) Choosing a rebuilder that out sources much of the work instead of doing it in house.
That should give you a good starting point.
Regardless of the standards; what matters most is how well the machine cuts metal!
Sounds like an interesting project. I would not mind being involved in some manner.
Hi. I also have been involved in machine rebuilding for over 45 years and live in Cottage Grove MN. I am currently rebuilding some machines in Crystal and if your in the Twin Cities I could come over and take a look. I know all the most of the rebuilders in the area. Some are excellent and some are bad. I also have taught machine building (and Rebuilding) in many new machine factories in Taiwan and Spinner a German CNC Lathe builder.
In Taiwan I consulted and taught at YCM and Quaser Machine who both use linear ways.
Richard King, King-Way Scraping Consultants, Home Page Richard@handscaping.com 651 338 8141
Thanks to the contributors.
Clive, I was involved in testing geometry correction software on 5-axis machines at K&T in the 1990's and wondered myself if some of the retrofit controls would provide this feature. It's not going to take out slop but could correct for unidirectional yaw... I see G&L (MAG) now offers this as an add-on (like much of the technology they got from K&T) and I'm sure it's not cheap.
Dinosaur, thank you for the article and list of standards; I am very familiar with ASME B5.54 as I was a "Machining Center" guy, so recently obtained the B5.57 spec for lathes. FWIW I've used NAS 979-1969 for 5-axis testing and circle-diamond-square (yes I've been called dinosaur as well).
Richard, would love to meet you. This is a grinding job as the underside of the back way need about .005 off the length (35 feet). I can't promise work, I'm sort of an internal consultant on this project but the money comes from my cohorts (rivals) at the company.
Hence the question remains: what is a reasonable linear accuracy & repeatability? Probably different X & Z values, I'm thinking X .0005 / .0003 radially, Z .005 / .001 (inches).
On new CNC and super precision machine tools we use the spec of .00005" per 12" and on conventional machine we say .0002" / 12". on the ways. The spec's are not always square as they are set to compensate for sag, tool pressure, weight and construction.
I just finished some Drake CNC Thread Grinders and on there spindle .00005" to .0001" run-out with a cold spindle. Many factories test the machines after they are warmed up. You need to look in the manual of the machine and see what the new machine spec's are. If it's not there, contact the Machine factory / Rep and ask them to see the new machine spec sheets. The spec's vary to the size of the machines.
Gurutzpe's USA office is in Minnesota.
The phone number below is to their service department.
Gurutzpe Machine Tools, Inc.
6616 Central Avenue NE
Minneapolis, MN 55432
Thanks for your earlier post Dinosaur, that's a very useful spec. Regards Tyrone.